Hand Gesture Recognition with Multiple Leap Motion Devices

Abstract

Gesture recognition is often used in various fields like sign language recognition, entertainment, virtual and augmented reality applications and many others. Vision based approaches with such devices like Leap Motion controllers and Kinect are often used in such systems and allow to achieve high performance in terms of recognition accuracy. This paper shows how one can overcome the most common problem of such systems - self-occlusion. In this paper we propose the method for the hand gesture recognition which requires 3 Leap Motion controller devices. This work shows that increasing the number of cameras can leads to better recognition accuracy. In addition we provide the methods of creating camera setup.

Keywords

• gesture recognition
• vr interfaces
• leap motion
• Human computer interaction (HCI)
• Tracking
• unity